Loading ...
Sorry, an error occurred while loading the content.

Re: [softrock40] Re: BS170 in Ensemble RXTX

Expand Messages
  • Alan
    ... Subject: [softrock40] Re: BS170 in Ensemble RXTX ... Steve, I would say it is not unexpected. The actual tolerances of the BS170 would probably give a
    Message 1 of 27 , May 1, 2011
    View Source
    • 0 Attachment
      ----- Original Message -----
      Subject: [softrock40] Re: BS170 in Ensemble RXTX



      > Just arrived at the final test stage and obtained the exact same
      > measurements across R48 and R49 previously reported by Paul (OE8PCK, 37mV
      > and 65mV). I have ordered a replacement BS 170.
      >
      > Is this a common problem?
      >

      Steve,

      I would say it is not unexpected.
      The actual tolerances of the BS170 would probably give a possible range of
      even more than that.
      I do not know when it actually becomes a problem. It seems reasonable that
      they should be matched.
      I wonder what difference it actually makes?
      There must be other factors to be considered if they are to be considered
      matched?

      Graham, Robby's figures are often just taken from the actual one he built,
      they are not necessarily typical.

      73 Alan G4ZFQ
    • Sid Boyce
      ... Mine are 40.6mV and 32.1mV respectively but I have low gain 2x the voltage at the gates, so I m going to change them, hoping that from a batch of 5 from RS
      Message 2 of 27 , May 1, 2011
      View Source
      • 0 Attachment
        On 01/05/11 16:55, Alan wrote:
        >
        > ----- Original Message -----
        > Subject: [softrock40] Re: BS170 in Ensemble RXTX
        >
        > > Just arrived at the final test stage and obtained the exact same
        > > measurements across R48 and R49 previously reported by Paul (OE8PCK,
        > 37mV
        > > and 65mV). I have ordered a replacement BS 170.
        > >
        > > Is this a common problem?
        > >
        >
        > Steve,
        >
        > I would say it is not unexpected.
        > The actual tolerances of the BS170 would probably give a possible range of
        > even more than that.
        > I do not know when it actually becomes a problem. It seems reasonable that
        > they should be matched.
        > I wonder what difference it actually makes?
        > There must be other factors to be considered if they are to be considered
        > matched?
        >
        > Graham, Robby's figures are often just taken from the actual one he built,
        > they are not necessarily typical.
        >
        > 73 Alan G4ZFQ
        >
        Mine are 40.6mV and 32.1mV respectively but I have low gain 2x the
        voltage at the gates, so I'm going to change them, hoping that from a
        batch of 5 from RS they will be fairly well matched.
        73 ... Sid.
        >


        --
        Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
        Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support
        Senior Staff Specialist, Cricket Coach
        Microsoft Windows Free Zone - Linux used for all Computing Tasks
      • grahamg3zod
        Similar to my results but a little more extreme. Both my BS170s came off a strip of three, by the way. It d be interesting to know the significance of the
        Message 3 of 27 , May 1, 2011
        View Source
        • 0 Attachment

          Similar to my results but a little more extreme.  Both my BS170s came off a strip of three, by the way.  It'd be interesting to know the significance of the differences - different gains or is it just a case of the bias voltages really need to be different/adjustable for the two transistors?

          73 de Graham G3ZOD FISTS #8385
          http://www.fists.co.uk


          --- In softrock40@yahoogroups.com, Sid Boyce <sboyce@...> wrote:
          > Mine are 40.6mV and 32.1mV respectively but I have low gain 2x the
          > voltage at the gates, so I'm going to change them, hoping that from a
          > batch of 5 from RS they will be fairly well matched.
          > 73 ... Sid.

        • Sid Boyce
          I had been wondering the same, but varying one may cause the other to vary also and you end up with driving one into saturation. The other idea would be to
          Message 4 of 27 , May 1, 2011
          View Source
          • 0 Attachment
            I had been wondering the same, but varying one may cause the other to
            vary also and you end up with driving one into saturation.
            The other idea would be to feed the secondary of T3 through 0.1 caps to
            the gates and apply separate bias voltages to the gates, but if the 2
            FET's are not nearly matched same situation could occur.
            73 ... Sid.

            On 01/05/11 18:02, grahamg3zod wrote:
            > Similar to my results but a little more extreme. Both my BS170s came off
            > a strip of three, by the way. It'd be interesting to know the
            > significance of the differences - different gains or is it just a case
            > of the bias voltages really need to be different/adjustable for the two
            > transistors?
            >
            > 73 de Graham G3ZOD FISTS #8385
            > http://www.fists.co.uk
            >
            >
            > --- In softrock40@yahoogroups.com, Sid Boyce <sboyce@...> wrote:
            > > Mine are 40.6mV and 32.1mV respectively but I have low gain 2x the
            > > voltage at the gates, so I'm going to change them, hoping that from a
            > > batch of 5 from RS they will be fairly well matched.
            > > 73 ... Sid.
            >
            >


            --
            Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
            Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support
            Senior Staff Specialist, Cricket Coach
            Microsoft Windows Free Zone - Linux used for all Computing Tasks
          • Sid Boyce
            I replaced both BS170 s in the PA stage and I have lots of RF out. Driving it with a tone generator program under Linux and I get 20V peak at the antenna. R48
            Message 5 of 27 , May 1, 2011
            View Source
            • 0 Attachment
              I replaced both BS170's in the PA stage and I have lots of RF out.
              Driving it with a tone generator program under Linux and I get 20V peak
              at the antenna.
              R48 = 46.5mV and R49 = 42.51mV.
              The heatsink gets pretty warm now, before there was no warmth.

              The originals were from the kit and the new ones, which I guessed would
              probably be from the same batch, came as a sealed set from RS Components.
              73 ... Sid.

              On 01/05/11 20:27, Sid Boyce wrote:
              > I had been wondering the same, but varying one may cause the other to
              > vary also and you end up with driving one into saturation.
              > The other idea would be to feed the secondary of T3 through 0.1 caps to
              > the gates and apply separate bias voltages to the gates, but if the 2
              > FET's are not nearly matched same situation could occur.
              > 73 ... Sid.
              >
              > On 01/05/11 18:02, grahamg3zod wrote:
              >> Similar to my results but a little more extreme. Both my BS170s came off
              >> a strip of three, by the way. It'd be interesting to know the
              >> significance of the differences - different gains or is it just a case
              >> of the bias voltages really need to be different/adjustable for the two
              >> transistors?
              >>
              >> 73 de Graham G3ZOD FISTS #8385
              >> http://www.fists.co.uk
              >>
              >>
              >> --- In softrock40@yahoogroups.com, Sid Boyce<sboyce@...> wrote:
              >> > Mine are 40.6mV and 32.1mV respectively but I have low gain 2x the
              >> > voltage at the gates, so I'm going to change them, hoping that from a
              >> > batch of 5 from RS they will be fairly well matched.
              >> > 73 ... Sid.
              >>
              >>
              >
              >


              --
              Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
              Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support
              Senior Staff Specialist, Cricket Coach
              Microsoft Windows Free Zone - Linux used for all Computing Tasks
            • grahamg3zod
              Sid, great news! For comparison, I wonder if you have a measurement of the RF out before you replaced them? Interestingly, one of your new readings is still
              Message 6 of 27 , May 2, 2011
              View Source
              • 0 Attachment

                Sid, great news!  For comparison, I wonder if you have a measurement of the RF out before you replaced them?

                Interestingly, one of your new readings is still out of spec. as far as the Builders' Guide is concerned - 42.51mV is below 45mV (55mV +/-10mV).  I know someone will say that the guide values are only approximate, but that is why the range is included and perhaps the 'Guide needs to make this broader, say +/- 15mV?

                My heatsink gets warm and I am measuring 1 watt out, but I'm driving it fairly hard with the audio from Rocky almost full on and the Delta 44 output level set to +4dBu (it's maximum setting).  Still a bit undecided whether to leave it alone or not.

                73 de Graham G3ZOD FISTS #8385
                http://www.fists.co.uk

                --- In softrock40@yahoogroups.com, Sid Boyce <sboyce@...> wrote:
                > I replaced both BS170's in the PA stage and I have lots of RF out.
                > Driving it with a tone generator program under Linux and I get 20V peak
                > at the antenna.
                > R48 = 46.5mV and R49 = 42.51mV.
                > The heatsink gets pretty warm now, before there was no warmth.
                >
                > The originals were from the kit and the new ones, which I guessed would
                > probably be from the same batch, came as a sealed set from RS Components.
                > 73 ... Sid.

              • Alan
                ... Subject: [softrock40] Re: BS170 in Ensemble RXTX - SUCCESS! ... the Builders Guide is concerned - 42.51mV is below 45mV (55mV +/-10mV). I know someone
                Message 7 of 27 , May 2, 2011
                View Source
                • 0 Attachment
                  ----- Original Message -----

                  Subject: [softrock40] Re: BS170 in Ensemble RXTX - SUCCESS!




                  >Interestingly, one of your new readings is still out of spec. as far as
                  the Builders' Guide is concerned - 42.51mV is below 45mV (55mV +/-10mV).
                  I know someone will say that the guide values are only approximate, but
                  that is why the range is included and perhaps the 'Guide needs to make
                  this broader, say +/- 15mV?

                  Graham,

                  We need someone to actually give a proper analysis of this.
                  I'm sure Robby will tell you he is no more of an electronics expert than I
                  am. His skill is in producing construction instructions. It is difficult
                  with a project like this to give absolute tolerances. Some readings may be
                  expected to be very close to nominal, others like PA current, can vary a
                  lot.
                  The general consensus is that a wider spread is acceptable, but how much?
                  If it really is critical perhaps the best match of the three should be
                  selected before construction.


                  >My heatsink gets warm and I am measuring 1 watt out, but I'm driving it
                  fairly hard with the audio from Rocky almost full on and the Delta 44
                  output level set to +4dBu (it's maximum setting). Still a bit undecided
                  whether to leave it alone or not.

                  It seems to me the rough and ready criteria is the same as any PA. Run it at
                  least a couple of dB less power than you get when you force it.
                  My Delta 44 will drive a Softrock to nearly one watt when set to the
                  "Consumer" level. That is with a 12V4 PSU.
                  Maybe I'm too careful but I keep mine at 1/2 watt. One watt makes far more
                  difference to the heatsink temperature than it does to the signal at the
                  other end.
                  However, many claim to use it at 1 watt or even more.

                  73 Alan G4ZFQ
                • Sid Boyce
                  I know 42.51mV (static reading with PTT on) is below spec, but I m getting RF out at a decent level now. I am not too keen to go through endless replacements.
                  Message 8 of 27 , May 2, 2011
                  View Source
                  • 0 Attachment
                    I know 42.51mV (static reading with PTT on) is below spec, but I'm
                    getting RF out at a decent level now. I am not too keen to go through
                    endless replacements. I'm thinking of using a jig to match up a pair.
                    73 ... Sid.

                    Before replacement the RF out was only around 4V p-p.
                    On 02/05/11 11:34, grahamg3zod wrote:
                    > Sid, great news! For comparison, I wonder if you have a measurement of
                    > the RF out before you replaced them?
                    >
                    > Interestingly, one of your new readings is still out of spec. as far as
                    > the Builders' Guide is concerned - 42.51mV is below 45mV (55mV +/-10mV).
                    > I know someone will say that the guide values are only approximate, but
                    > that is why the range is included and perhaps the 'Guide needs to make
                    > this broader, say +/- 15mV?
                    >
                    > My heatsink gets warm and I am measuring 1 watt out, but I'm driving it
                    > fairly hard with the audio from Rocky almost full on and the Delta 44
                    > output level set to +4dBu (it's maximum setting). Still a bit undecided
                    > whether to leave it alone or not.
                    >
                    > 73 de Graham G3ZOD FISTS #8385
                    > http://www.fists.co.uk
                    >
                    > --- In softrock40@yahoogroups.com, Sid Boyce <sboyce@...> wrote:
                    > > I replaced both BS170's in the PA stage and I have lots of RF out.
                    > > Driving it with a tone generator program under Linux and I get 20V peak
                    > > at the antenna.
                    > > R48 = 46.5mV and R49 = 42.51mV.
                    > > The heatsink gets pretty warm now, before there was no warmth.
                    > >
                    > > The originals were from the kit and the new ones, which I guessed would
                    > > probably be from the same batch, came as a sealed set from RS Components.
                    > > 73 ... Sid.
                    >
                    >


                    --
                    Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
                    Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support
                    Senior Staff Specialist, Cricket Coach
                    Microsoft Windows Free Zone - Linux used for all Computing Tasks
                  • grahamg3zod
                    Hi Alan and thanks for the comments. Which band(s) were you using with the audio settings you describe? I did try looking at a BS170 data sheet, but as my
                    Message 9 of 27 , May 3, 2011
                    View Source
                    • 0 Attachment

                      Hi Alan and thanks for the comments.  Which band(s) were you using with the audio settings you describe?

                      I did try looking at a BS170 data sheet, but as my knowledge of electronics is limited, I found it fairly tough to understand, other than to note that some parameters have wide variations noted in the sheet.

                      With the audio settings full on, I was getting a bit more than 1.5 watts on 15m, so I did back it off to 1 watt.  I tried it out on 12m and 10m yesterday and was getting, with the same audio settings, 0.8 watt on 12m and 0.6 watt on 10m, so all looks satisfactory to me despite the (apparently) low drain/source standing current on the BS170's.

                      The previous comments about the heat sink just getting warm were misleading - touching the surface of of the vertical fins feels warm but putting a finger against the side of the heat sink revealed that it was actually pretty hot rather than warm (this is using CW rather than 100% duty cycle).

                      I think I'll leave the transceiver as-is now unless someone comes up with a good reason to find PA transistors that run with a higher drain/source current.

                      73 de Graham G3ZOD FISTS #8385
                      http://www.fists.co.uk


                      --- In softrock40@yahoogroups.com, "Alan" <alan4alan@...> wrote:
                      > We need someone to actually give a proper analysis of this.
                      > I'm sure Robby will tell you he is no more of an electronics expert than I
                      > am. His skill is in producing construction instructions. It is difficult
                      > with a project like this to give absolute tolerances. Some readings may be
                      > expected to be very close to nominal, others like PA current, can vary a
                      > lot.
                      > The general consensus is that a wider spread is acceptable, but how much?
                      > If it really is critical perhaps the best match of the three should be
                      > selected before construction.
                      > ...
                      > It seems to me the rough and ready criteria is the same as any PA. Run it at
                      > least a couple of dB less power than you get when you force it.
                      > My Delta 44 will drive a Softrock to nearly one watt when set to the
                      > "Consumer" level. That is with a 12V4 PSU.
                      > Maybe I'm too careful but I keep mine at 1/2 watt. One watt makes far more
                      > difference to the heatsink temperature than it does to the signal at the
                      > other end.
                      > However, many claim to use it at 1 watt or even more.
                      >
                      > 73 Alan G4ZFQ

                    • Alan
                      ... Subject: [softrock40] Re: BS170 in Ensemble RXTX - SUCCESS! ... the audio settings you describe? Graham, To elaborate. I have 3 RXTXs a 6.1 80/60/40m. 6.3
                      Message 10 of 27 , May 3, 2011
                      View Source
                      • 0 Attachment
                        ----- Original Message -----
                        Subject: [softrock40] Re: BS170 in Ensemble RXTX - SUCCESS!



                        >Which band(s) were you using with
                        the audio settings you describe?

                        Graham,

                        To elaborate.
                        I have 3 RXTXs a 6.1 80/60/40m. 6.3 160-10m. Ensemble 30/20/17m.
                        I believe all have the same circuit components and values.
                        I now run the Delta 44 at -4dBu and attenuate/reduce gain so each will
                        deliver almost 1 watt maximum.
                        Each step of the D44 control gives 6dB less power. I used this setup so I
                        could easily use flea-power with WSPR using the TX audio adjustment and/or
                        the D44 steps.

                        The 6.1 has an extra 33K in the opamp input = 43K. I reckon that makes the
                        op-amp gain 0.23, power ~ -13dB.
                        The 6.3 has a crude attenuator/gain reduction which I reckon gives it -8dB.
                        This Softrock gives at least 500mW except on 10m. (400mW max) I'm not sure
                        if more audio will increase 10m power. I run it at 200mW, waiting for REAL
                        sunspots.
                        The Ensemble has an extra 10K in the opamp input = 20K. I reckon that
                        makes the op-amp gain 0.5, power ~-6dB

                        >I did try looking at a BS170 data sheet, but as my knowledge of
                        electronics is limited, I found it fairly tough to understand, other
                        than to note that some parameters have wide variations noted in the
                        sheet.

                        In all the years I have played radio I have avoided the mathematical and
                        theoretical side as much as possible.
                        My guess is that the Gate Threshold Voltage range of 0.8 to 3V for 1mA will
                        be reflected at higher voltages, leading to the differences we see.
                        But if the Gate Threshold Voltage of a device was 3V it seems that no
                        current would flow but it would still be in spec??

                        73 Alan G4ZFQ
                      • taka_sugi1206
                        Hello Paul. I had also same problem, but not so large difference as yours. The difference is very important and affects to TX IMD, output power. I suggest to
                        Message 11 of 27 , May 5, 2011
                        View Source
                        • 0 Attachment
                          Hello Paul.
                          I had also same problem, but not so large difference as yours.
                          The difference is very important and affects to TX IMD, output power.

                          I suggest to buy many BS170s and select almost same Is(FET source current) BS170s by using simple test circuit as follows.

                          Vd=12v, Rd(Drain load resistor)=1kohm, Rs(source resistor)=2.2ohm,
                          Vg(gate bias)=2.2 ¨C 2.5v, and then measure the voltage of Rs.

                          For my RXTX, voltage of R48,R49 is 54mv(24.5mA),53mv(24.1mA)each, and
                          max output power is about 1w, 3 rd order IMD : about 25dB£À0.7w output

                          73 de Taka, JA1TLH
                        • Adam Sampson
                          ... It might be worth having a guide to matching BS170s in the build instructions. The BS170s in my RXTX aren t very well matched, and they re cheap enough
                          Message 12 of 27 , May 6, 2011
                          View Source
                          • 0 Attachment
                            Sid Boyce <sboyce@...> writes:

                            > I'm thinking of using a jig to match up a pair.

                            It might be worth having a guide to matching BS170s in the build
                            instructions. The BS170s in my RXTX aren't very well matched, and
                            they're cheap enough (and useful enough for other purposes) that buying
                            a couple of dozen to sort through is practical.

                            --
                            Adam Sampson <ats@...> <http://offog.org/>
                          • Oliver Goldenstein
                            ... thanks for that circuit Taka! 73 de Oliver, DL6KBG -- Oliver, DL6KBG, JO61UB Blog: http://dl6kbg.blogspot.com
                            Message 13 of 27 , May 6, 2011
                            View Source
                            • 0 Attachment
                              > I suggest to buy many BS170s and select almost same Is(FET source current) BS170s  by  using  simple test circuit as follows.
                              >
                              >  Vd=12v, Rd(Drain load resistor)=1kohm, Rs(source resistor)=2.2ohm,
                              >  Vg(gate bias)=2.2 ¨C 2.5v,  and then measure the voltage of Rs.

                              thanks for that circuit Taka!

                              73 de Oliver, DL6KBG

                              --
                              Oliver, DL6KBG, JO61UB
                              Blog: http://dl6kbg.blogspot.com
                            • Sid Boyce
                              ... I have 8 more around, just a matter of when I can get around to matching up a pair. 73 ... Sid. -- Sid Boyce ... Hamradio License G3VBV, Licensed Private
                              Message 14 of 27 , May 6, 2011
                              View Source
                              • 0 Attachment
                                On 06/05/11 13:15, Adam Sampson wrote:
                                > Sid Boyce <sboyce@... <mailto:sboyce%40blueyonder.co.uk>>
                                > writes:
                                >
                                > > I'm thinking of using a jig to match up a pair.
                                >
                                > It might be worth having a guide to matching BS170s in the build
                                > instructions. The BS170s in my RXTX aren't very well matched, and
                                > they're cheap enough (and useful enough for other purposes) that buying
                                > a couple of dozen to sort through is practical.
                                >
                                > --
                                > Adam Sampson <ats@... <mailto:ats%40offog.org>> <http://offog.org/>
                                >

                                I have 8 more around, just a matter of when I can get around to matching
                                up a pair.
                                73 ... Sid.
                                --
                                Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
                                Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support
                                Senior Staff Specialist, Cricket Coach
                                Microsoft Windows Free Zone - Linux used for all Computing Tasks
                              • taka_sugi1206
                                Hello, Oliver Sorry, there was one mistake... Rd= 1kohm is too large, pse change to 100 ohm. Please note heat of the FET . I bought many BS170s of
                                Message 15 of 27 , May 8, 2011
                                View Source
                                • 0 Attachment
                                  Hello, Oliver

                                  Sorry, there was one mistake...
                                  Rd= 1kohm is too large, pse change to 100 ohm.
                                  Please note "heat of the FET".

                                  I bought many BS170s of Fairchild,Mot, On-Semi and selected BS170‚"
                                  by using this simple test circuit.

                                  73@de Taka/JA1TLH


                                  > > I suggest to buy many BS170s and select almost same Is(FET source current) BS170s  by  using  simple test circuit as follows.
                                  > >
                                  > >  Vd=12v, Rd(Drain load resistor)=1kohm, Rs(source resistor)=2.2ohm,
                                  > >  Vg(gate bias)=2.2 ¨C 2.5v,  and then measure the voltage of Rs.
                                  >
                                  > thanks for that circuit Taka!
                                  >
                                  > 73 de Oliver, DL6KBG
                                  >
                                  > --
                                  > Oliver, DL6KBG, JO61UB
                                  > Blog: http://dl6kbg.blogspot.com
                                  >
                                • Hugh
                                  ... Me too. I am building up RxTx kits numbers 2 and 3, and decided to test all 18 BS170s Tony sent to me (after finding considerable inbalance in the first
                                  Message 16 of 27 , Jul 23, 2011
                                  View Source
                                  • 0 Attachment
                                    > > Q8 and Q7 appear to have large differences...
                                    >
                                    > Yes, this has worried several builders.

                                    Me too.

                                    I am building up RxTx kits numbers 2 and 3, and decided to test all 18 BS170s Tony sent to me (after finding considerable inbalance in the first kit I built.. well outside the construction notes)

                                    Nature of test: put +2.20V on the gate, drain with 1KOhm in series, Source with 2.2Ohms and measure the source current.

                                    I know this basically only measures the gate turn-on, but this is probably the most important for the matching between the PA pair and the bias transistor, as both PA transistors share a common bias voltage in the circuit.
                                    Results: (mA) (in order as they appear on the stips)
                                    25, 25, 37, 14, 25, 27, 12, 14, 9, 26, 9, 16, 39, 39, 23, 14

                                    OK, the test is rough and ready, but I there is quite some variation (factor of 4 in current) (I will not be using the 9mA device)
                                    Devices were fresh from the silver-strip and ESD precations were observed. I don't know if something 'nasty' happend between USA and Europe in shipping, but I doubt it. I will be using the 3x 25mA and the 3x39 for the Pa/Bias 3 transistors. Thanks Tony for supplying extra! I was also little confused as each kit comes as 2 seperately mailed jiffy-bags!

                                    In hindsight, I don't think there is much effect on the transmitter, as the first one I made seems OK and the harmonics are OK, but I think it is always best to balance out the pair as best as possible to minimise harmonic generation, and get the bias point about right for the correct distortion specs.
                                    I guess when I have built the transmitter and keyed it up, I will discover the other transistor parameter mismatch.. (gm vatiation).
                                    Hope this has been helpful, Happy building!
                                  Your message has been successfully submitted and would be delivered to recipients shortly.